This project proposes to develop a robust prognostic model for patients with high-risk primary breast cancer (HRPBC). We will use as a base our clinical prognostic model, which was developed and validated in HRPBC patients treated with the high-dose regimen STAMP-I. This model assigns each patient a score and segregates two groups with either high or low risk of relapse after high-dose chemotherapy (HDC). We will test this model in large data sets of HRPBC patients treated with standard-dose chemotherapy (SDC), such as those enrolled in the CALGB (Cancer And Leukemia Group B) trial 8541, or in several prospective SWOG (South Western Oncology Group) randomized studies. In addition, we will validate the model in groups of patients treated with other HDC regimens, such as: 1) patients registered at IBMTR (International Bone Marrow Transplant Registry)/ABMTR (Autologous Bone Marrow Transplant Registry); and 2) patients accrued in the Dutch randomized trial of SDC vs. HDC in patients with four or greater involved nodes. Additionally, we intend to study the following angiogenesis markers in tumor blocks from 206 HRPBC patients treated at our program with HDC, and followed for a median of five years: microvessel density, vascular endothelial growth factor (VEGF), and alpha-v-beta-3 integrin. The driving hypothesis is that a highly angiogenic phenotype contributes to the poor outcome observed in patients with a high clinical score, which HDC does not appear to overcome, as suggested by our preliminary observations that the major impact of HDC may be restricted to the group of patients with a low clinical score. Proof of this hypothesis would offer the impetus for subsequent trials testing combinations of novel angiogenesis-targeting agents with HDC, which is capable of massively cytoreducing the tumor. The rationale for such combinations is based on the synergy observed when these new agents are administered concurrently with chemo or radiotherapy, and on the opportunity to test these agents against post-transplant minimal residual disease, when its effectiveness might be higher than against bulky tumors.